Despite 30 years of study, gaseous elemental mercury (Hg(0)) exchange magnitude and controls between terrestrial surfaces and the atmosphere still remain uncertain. We compiled data from 132 studies, including 1290 reported fluxes from more than 200,000 individual measurements, into a database to statistically examine flux magnitudes and controls. We found that fluxes were unevenly distributed, both spatially and temporally, with strong biases toward Hg-enriched sites, daytime and summertime measurements. Fluxes at Hg-enriched sites were positively correlated with substrate concentrations, but this was absent at background sites. Median fluxes over litter- and snow-covered soils were lower than over bare soils, and chamber measurements showed higher emission compared to micrometeorological measurements. Due to low spatial extent, estimated emissions from Hg-enriched areas (217 Mg·a(-1)) were lower than previous estimates. Globally, areas with enhanced atmospheric Hg(0) levels (particularly East Asia) showed an emerging importance of Hg(0) emissions accounting for half of the total global emissions estimated at 607 Mg·a(-1), although with a large uncertainty range (-513 to 1353 Mg·a(-1) [range of 37.5th and 62.5th percentiles]). The largest uncertainties in Hg(0) fluxes stem from forests (-513 to 1353 Mg·a(-1) [range of 37.5th and 62.5th percentiles]), largely driven by a shortage of whole-ecosystem fluxes and uncertain contributions of leaf-atmosphere exchanges, questioning to what degree ecosystems are net sinks or sources of atmospheric Hg(0).